calibre 0.6 betas now available

[tselling]

I am still seeing a problem with bulk converting alot of files. Calibre dies on me.

Also I recently got a Jetbook and tried using my mobipocket files on it. I am seeing a problem with some of the mobipocket ebooks with quotes and apostrophes showing up as strange characters. After doing some searces and experimentation I fixed the problem in one book by reconverting it with input character encoding set to utf-8. It seems like I shouldn't have to set it explicitly. I am unsure if this is a calibre conversion problem since not all my mobipocket books have this problem.

[kovidgoyal]

@ralob: The only way I can think of is setting up remote desktop, but it will have to wait a few days as I'm rather busy at the moment

@tselling: If your input documents don't specify a correct character encoding, you have to specify it manually in the conversion options. As for the crasing on the Bulk convert, that's another of the cant replicate issues, so it's hard for me to debug. Just Bulk convert in smaler batches for now.

[richreads]

Hi.
I'm getting the following when i try and convert.
it has happened converting txt, htm, lit and pdb.

Creating LRF Output...
Traceback (most recent call last):
File "worker.py", line 103, in <module>
File "worker.py", line 90, in main
File "calibre\gui2\convert\gui_conversion.pyo", line 17, in gui_convert
File "calibre\ebooks\conversion\plumber.pyo", line 721, in run
File "calibre\ebooks\lrf\output.pyo", line 179, in convert
File "calibre\ebooks\lrf\html\convert_from.pyo", line 1894, in process_file
File "re.pyo", line 150, in sub
TypeError: expected string or buffer

[Landrovan]

Hi,

Just curious

I just look a ticket #2909. You say that you create 3 calibre-parallel process by core. Where do the 3 come? Does that mean when doing a lot of job calibre do you allow 3 job on each processor or those that mean you need 3 process by job? Or do I'm way off.

Thanks

[kovidgoyal]

@richreads: Should be fixed in next release, though I can't duplicate that error.

@Landrovan; calibre runs only 1 process per processor. The reason there are 3 is because one of those three is used for conversion jobs, one for adding of books and one for saving of books.

[=X=]

I tried to log this in the database but I get an Error login on, did you delete user accounts on your defect tracking system? <<<EDIT: Never mind it looks like you did nuke old accounts I was able to recreate an account with my same username>>>>

DESC:
Regardless of what PDF I try to crop I always get the following error

pdfmanipulate crop mypdf.pdf

I've even ran the tool using the following commands

pdfmanipulate.exe crop -o out.pdf  -x 120 -y 724 -v 560 -w 70 mybig.pdf

Traceback (most recent call last):
File "cli.py", line 72, in <module>
File "cli.py", line 61, in main
File "calibre\ebooks\pdf\manipulate\crop.pyo", line 144, in main
File "calibre\ebooks\pdf\manipulate\crop.pyo", line 114, in crop_pdf
TypeError: unsupported operand type(s) for -: 'NumberObject' and 'unicode'

[cszhy]

Hi I want to hotsync the articles in my google reader to prs500 ,
is there any receipt I can use in calibre
thanks

[tselling]

[QUOTE=kovidgoyal;530466@tselling: If your input documents don't specify a correct character encoding, you have to specify it manually in the conversion options. As for the crasing on the Bulk convert, that's another of the cant replicate issues, so it's hard for me to debug. Just Bulk convert in smaler batches for now.[/QUOTE]

Regarding the character encoding problem: I have found that EPub ebooks on the Jetbook do not have the character encoding problems. Some of the mobipocket ebooks were converted from original epub ebooks... and some of the EPub ebooks were converted from original mobipocket format. Is it possible that character encoding got lost in the conversion? I am not sure why none of the EPub ebooks seem to have the problem... does character encoding work differently for EPub books?

Regarding the crashing, I have around 150 ebooks in my Calibre library running on Vista 64. If I try to convert all 150 books (say mobi to mobi conversion to add UTF-8 character encoding), Calibre always crashes (Not responding error where I have to click OK and the program quits). If I convert 25 books at a time it usually works ok, but may crash before the last 25 books are converted. If I convert 50 books at a time ususally works ok the first time but often crashes on the next 50, or the last 50.

PS. Thanks for the great work on this application. Its a great tool!

[itimpi]

I am finding that Callibre RAM usage seems rather excessive!

I had Calibre running. It was idling using about 125Mb (for about 6400 books in the library). I then selected 240 books for conversion. I watched the memory usage of the main Calibre process grow to about 1.25GB. During this whole time the GUI was locked out, and none of the (3) Callibre-Parallel.exe processes appeared to be active. I assume this was Calibre queuing the books for conversion? At the end of this process the GUI came back and one of the Calibre-Parallel.exe processes became active and I assume is now doing the actual conversions.

The conversions are now running. The main Calibre process has not yet released any of the RAM it has grabbed after about 60 of the conversions have completed. I will have to wait and see if it does it at the end of all of the conversions.

This sort of RAM uage might explain why some users are experiencing crashes trying to convert a lot of books. What is Calibre doing that needs that sort of memory? Is there anything I can do to try and determine what it is that is consuming that amount of RAM? Is it possible there is a memory leak?

In case it matters I am running Windows XP Pro+SP3 with a 2.6GHz single Core system and 2GB of RAM. I was doing a bulk conversion to ePub.

EDIT: Calibre crashed before completing the conversions so, I was not able to determine if a clean finish would have released the RAM.

[MovieHQ]

I am getting the following error when I plug my 505 in. It works fine with the non-beta version. I am on Vista 32bit.

unsupported operand type(s) for +: 'NoneType' and 'str'

Traceback (most recent call last):
File "calibre\gui2\device.pyo", line 57, in run
File "calibre\gui2\device.pyo", line 191, in _books
File "calibre\devices\prs505\driver.pyo", line 96, in books
TypeError: unsupported operand type(s) for +: 'NoneType' and 'str'

Any ideas?

[kovidgoyal]

@MovieHQ: Do you have a card in your reader?

@itimpi: Certainly sounds like a memory leak, the mem usage of the main calibre process should not grow that much when queueing books for conversion. The relevant function where the queueing happens is convert_bulk_ebooks in calibre.gui2.tools

Looking over the code, I don't see any obvious memory leaks and I don't have the time to debug it in detail right now. Maybe you will have better luck?

[itimpi]

Originally Posted by kovidgoyal

@itimpi: Certainly sounds like a memory leak, the mem usage of the main calibre process should not grow that much when queueing books for conversion. The relevant function where the queueing happens is convert_bulk_ebooks in calibre.gui2.tools

Looking over the code, I don't see any obvious memory leaks and I don't have the time to debug it in detail right now. Maybe you will have better luck?

Thought I should let you know that the RAM is NOT released when the conversions finished just in case it triggers a thought from your end as to where the RAM might be getting allocated.

I will try and have a look at the source tomorrow to see if I can spot anything that might cause this behavior.

I wonder if anyone else has spotted this behavior and whether it is Windows specific or common to all platforms? If it happens on all platforms it will probably be much easier to find.

[sapient]

Conversion from a zipped html book is problematic. I tried this:
http://www.freenetpages.co.uk/hp/alan.gauld/tutor.zip
It mixed up the order of the pages. Instead of appending the different html files in alphabetical (?) order, it should parse index.html and follow the order specified therein.

[MovieHQ]

Yes, I do have a card in the reader.

[kovidgoyal]

@sapient: index.htm doesn't specify the order in your example tutcont.htm does. Unzip the zip file andd add tutcont.htm to calibre

@MoviHQ: Does it work without the card?